Cellular and Molecular Neurobiology https://doi.org/10.1007/s10571-018-0614-5

REVIEW PAPER

Resetting the Stress System with a Mifepristone Challenge

Sergiu Dalm1 · Adriaan M. Karssen1 · Onno C. Meijer1,2 · Joseph K. Belanoff3 · E. Ronald de Kloet1,2

Received: 15 July 2018 / Accepted: 18 August 2018 © The Author(s) 2018

Abstract Psychotic depression is characterized by elevated circulating cortisol, and high daily doses of the glucocorticoid/progesterone antagonist mifepristone for 1 week are required for significant improvement. Using a rodent model, we find that such high doses of mifepristone are needed because the antagonist is rapidly degraded and poorly penetrates the blood–brain barrier, but seems to facilitate the entry of cortisol. We also report that in male C57BL/6J mice, after a 7-day treatment with a high dose of mifepristone, basal blood corticosterone levels were similar to that of vehicle controls. This is surprising because after the first mifepristone challenge, corticosterone remained elevated for about 16 h, and then decreased towards vehicle control levels at 24 h. At that time, stress-induced corticosterone levels of the 1xMIF were sevenfold higher than the 7xMIF group, the latter response being twofold lower than controls. The 1xMIF mice showed behavioral hyperactivity during exploration of the circular hole board, while the 7xMIF mice rather engaged in serial search patterns. To explain this rapid reset of corticosterone secretion upon recurrent mifepristone administration, we suggest the following: (i) A rebound gluco- corticoid feedback after cessation of mifepristone treatment. (ii) Glucocorticoid agonism in transrepression and recruitment of cell-specific coregulator cocktails. (iii) A more prominent role of brain MR function in control of stress circuit activity. An overview table of neuroendocrine MIF effects is provided. The data are of interest for understanding the mechanistic underpinning of stress system reset as treatment strategy for stress-related diseases.

Keywords Stress · Brain · Behavior · RU38486 · Glucocorticoid receptor · Mineralocorticoid receptor

Introduction restores aberrant levels of the corticosteroids (Murphy et al. 1993; Belanoff et al. 2001, 2002; DeBattista and Belanoff Patients suffering from psychotic major depression benefit 2006; Flores et al. 2006; Blasey et al. 2009, 2011; Block from a brief treatment with the glucocorticoid/progester- et al. 2018), see for meta-analysis (Garner et al. 2016). The one receptor antagonist RU38486 or mifepristone (MIF), fast amelioration of psychotic and depressive symptoms is in a dose range of 600–1200 mg/day, once a day for four to thought to be at least in part due to restoration of glucocor- seven days. This high dose of the antiglucocorticoid rapidly ticoid action to which the untreated patient with psychotic improves emotional expressions and cognitive abilities, and depression is resistant, while the anti-progestin activity of MIF is not implicated (Belanoff et al. 2001; Thomson and Craighead 2008). A recent analysis of all controlled phase 2 Sergiu Dalm and Adriaan M. Karssen have contributed equally to and 3 studies (MIF: n = 833 and placebo: n = 627) suggested this work. that a dose of 1200 mg MIF/day for 7 days significantly * E. Ronald de Kloet reduced psychotic symptoms. For an effective treatment, cir- [email protected]; [email protected] culating plasma levels of ≥ 1637 ng MIF/ml blood appeared required, while under such conditions basal HPA-axis activ- 1 Division of Medical Pharmacology, Leiden/Amsterdam Center for Drug Research and Leiden University Medical ity was increased (Block et al. 2017, 2018). The data raise Center, Leiden University, P.O. Box 9502, 2300 RA Leiden, the question how MIF can exert this rapid effect on psychotic The Netherlands major depression. 2 Department of Medicine, Division of Endocrinology, Leiden In the current study, we will address this question from University Medical Center, Room C‑7‑44, Postal zone C7‑Q, the perspective of adjustment in setpoint of stress system PO Box 9600, Leiden, The Netherlands activity by the antiglucocorticoid (Ratka et al. 1988; Hu 3 Corcept Therapeutics, Menlo Park, CA, USA

Vol.:(0123456789)1 3 Cellular and Molecular Neurobiology et al. 2012). Previous studies have shown that the effect Fifth, continuous infusion of 100 ng MIF/hr i.c.v. for 3 of MIF on the HPA axis is dependent on the dose, route, days enhanced stress-induced corticosterone secretion and and daily frequency of administration. In addition, a more increased the circadian amplitude of basal corticosterone prominent role of the brain mineralocorticoid receptors level (Van Haarst et al. 1996; van Haarst et al. 1997). Con- (MRs) after blockade of GR is likely, while also extrahypo- tinuous i.c.v. infusion improved spatial memory, which was thalamic circuits are involved. Table 1 presents the published impaired, however, when MIF was administered daily as MIF effects on the HPA axis, which can be summarized as a bolus of 100 ng/rat i.c.v. immediate before or after the follows. learning trial (Oitzl and de Kloet 1992; Oitzl et al. 1998). First, a central action of MIF is likely, since more than The findings reinforce the notion of temporal and contextual 100,000-fold lower dose centrally (10–100 ng local and diversity in glucocorticoid actions in various brain regions i.c.v.) than systemically (10–200 mg s.c. /kg rat) could (Joëls and de Kloet 1992; Joëls et al. 2018). achieve feedback blockade in behaviorally active fashion Finally, twice rather than once a day MIF enhanced (de Kloet et al. 1988; Ratka et al. 1989; van Haarst et al. HPA-axis activity in diabetic rats (Revsin et al. 2009; Stra- 1997; Dalm et al. 2008). In spite of rapid degradation and nahan et al. 2008) and, if applied at the end of three weeks poor brain penetration, such doses of MIF in the 10–200 mg daily stress exposure, blocked the reduction in neurogen- range can enter the brain and do translocate the immunoreac- esis (Oomen et al. 2007; Mayer et al. 2006); even a single tive (ir) glucocorticoid receptor (GR) to the neuronal nuclei day treatment on day 18 was effective in this paradigm (Hu (Van Eekelen et al. 1987; de Kloet 1991). This finding was et al. 2012). MIF administered twice a day, at postnatal days confirmed and further extended by the association of the 26–28, normalized deficits in hippocampal-dependent cog- MIF–GR complex to hippocampal Glucocorticoid Response nitive functions and associated neuronal activity that were Elements (GRE’s) (Spiga et al. 2011). 3H-MIF was found to previously induced by early-life maternal deprivation of rats bind to glucocorticoid receptor sites in human brain slices (Arp et al. 2016; Loi et al. 2017). in vitro (Sarrieau et al. 1986). A striking aspect of MIF’s efficacy is the ability to Second, the prolonged and profound increase in secretion reset the stress system. This phenomenon was previously of corticosterone was the expected outcome of MIF interfer- observed when the adrenals were rapidly (within 30 s under ence with the stress-induced GR-mediated negative feedback anesthesia) removed immediately following a severe stress (Ratka et al. 1989). The low GR occupancy during basal to prevent the surge in glucocorticoid secretion. We noted a trough conditions explains the lack of acute MIF effects, long-lasting (> 1 week) and profound potentiation in mor- but the antagonist can block during the p.m. phase or after phine- or β-endorphin-induced analgesia, which was corre- minor stressors provided corticosterone levels are increased lated with increased opioid receptor binding and decreased (Reul and de Kloet 1985). hippocampal enkephalin and dynorphin mRNA expression. Third, following MIF, the initial stress-induced rise in The antinociceptive effect of morphine was normalized if HPA-axis activity was attenuated. This blunted HPA-axis corticosterone was administered at the time of adrenalec- response likely was due to activation of co-localized hip- tomy (ADX) to mimic the stress or circadian rise. Also, pocampal MRs. These MRs mediate rapid actions of corti- 100 ng MIF i.c.v. administered during the p.m. circadian costerone that precede those of GRs (de Kloet 1991; Joëls rise at 1 h prior to ADX caused long-term sensitization to and de Kloet 1992; Karst et al. 2005; Joëls 2006). Indeed, morphine, apparently because the impact of the high pre- central (hippocampal) MR blockade disinhibited the HPA- stress corticosterone in brain was antagonized (Ratka et al. axis activity in rodents (Ratka et al. 1989; Oitzl et al. 1995; 1988; Iglesias et al. 1991). van Haarst et al. 1997) and humans (Dodt et al. 1993; Young In the present study, we mimicked in mice the high-dose et al. 1998; Deuschle et al. 1998). regimen of MIF that was beneficial in the patient studies Fourth, repeated daily administration for 5 days up to (Block et al. 2018). For this purpose, we applied a non- three weeks revealed an unexpected apparent GR agonism of invasive stress-free method for steroid delivery via oats MIF and blocked basal and stress-induced HPA-axis activ- (Dalm et al. 2008). After the first (1xMIF) and the seventh ity in rodents (Havel et al. 1996; Wulsin et al. 2010; van der administration (7xMIF) we assessed (i) the circadian corti- Veen et al. 2013). Similar GR agonism rather than antag- costerone secretion pattern; (ii) the behavioral and corticos- onism was noted with the novel selective GR modulators terone response to novelty during exploration of a circular C-108297, C-125281, and the GR modulator/MR antagonist hole board at 24 h post-treatment. Hippocampal and hypo- C-118335 (Solomon et al. 2014; Nguyen et al. 2017; Kroon thalamic MR, GR, and CRH mRNA expressions were also et al. 2018; Van den Heuvel et al. 2016). Repeated MIF treat- measured. Then, we investigated metabolism and brain pen- ment caused a differential pattern of activation and inhibi- etration of MIF using Liquid chromatography–Mass Spec- tion of central inputs to the PVN, as judged from c-FOS trometry–Mass Spectrometry (LC/MS/MS). Monolayers activation (Wulsin et al. 2010). of pig kidney epithelial cells (LLC-PK1) stably expressing

1 3 Cellular and Molecular Neurobiology et al. ( 2018 ) et al. ( 2018 ) van der Veen et al. ( 2013 ) et al. der Veen van Kroon et al. ( 2018 ) et al. Kroon van den Heuvel et al. ( 2016 ) et al. den Heuvel van ), Kroon ( 2016 ), Kroon et al. den Heuvel van ), Nguyen et al. et al. ( 2017 ), Nguyen et al. Nguyen Solomon et al. ( 2014 ) Solomon et al. Wulsin et al. ( 2010 ) et al. Wulsin Havel et al. ( 1996 ) et al. Havel Dalm et al. ( 2008 ) Dalm et al. Van Haarst et al. ( 1997 ) et al. Haarst Van Van Haarst et al. ( 1997 ) et al. Haarst Van Van Haarst et al. ( 1996 ) et al. Haarst Van de Kloet et al. ( 1988 ) et al. de Kloet Ratka et al. ( 1989 ) et al. Ratka Ratka et al. ( 1989 ) et al. Ratka Ratka et al. ( 1989 ) et al. Ratka Ratka et al. ( 1989 ) et al. Ratka Reference - - induced corticosterone secretion activity, no effect on thymus and no effect on thymus activity, adrenals terone level, decreased adrenal and adrenal decreased level, terone weight thymus corticosterone levels, decreased decreased corticosterone levels, weight & thymus adrenal HPA-axis activity HPA-axis HPA-axis activity HPA-axis HPA-axis activity HPA-axis HPA-axis activity, thymus weight weight thymus activity, HPA-axis reduced corticosterone terone levels at 1 h levels terone terone levels at 1 h levels terone corticosterone levels axis & interfered with retention with retention axis & interfered immobility tion of stress-induced corticostertion of stress-induced one for 4 h for tion of stress-induced corticostertion of stress-induced one for 4 h for Suppressed basal and amphetamine- Suppressed Restores high-fat disturbed HPA-axis HPA-axis disturbed high-fat Restores Decreased basal am and pm corticosDecreased - Decreased basal am and pm (trend) basal am and pm (trend) Decreased Suppressed basal and stress-induced basal and stress-induced Suppressed Suppressed basal and stress-induced basal and stress-induced Suppressed Suppressed basal and stress-induced basal and stress-induced Suppressed Suppressed basal and stress-induced basal and stress-induced Suppressed Increased basal and stress-induced basal and stress-induced Increased Decreased basal ACTH and corticos - basal ACTH Decreased Increased basal ACTH and corticos - basal ACTH Increased Increased peak and decreased basal peak and decreased Increased Doses ≥ 10 ng disinhibited HPA - secre Blunted peak and prolonged No change in basal corticosterone change No - secre Blunted peak and prolonged No change in basal corticosterone change No Outcome before amphetamine before injection injection injection test oral infusion daily for 5 days, 2.5 h 5 days, for infusion daily oral orally, daily for 3 weeks in chow 3 weeks for daily orally, orally, daily for 3 weeks in chow 3 weeks for daily orally, orally, daily for 3 weeks in chow 3 weeks for daily orally, s.c. 5 days, daily, 90 min after last daily, s.c. 5 days, s.c. 5 days, daily, 90 min after last daily, s.c. 5 days, s.c. 5 days, daily, 90 min after last daily, s.c. 5 days, s.c., 2 weeks daily s.c., 2 weeks orally via oats orally bilateral dorsal hippocampus bilateral i.c.v 100 ng/h/for 3 days 100 ng/h/for i.c.v. immediate after initial swim i.c.v. i.c.v. 15 min prior stressor to i.c.v. i.c.v. morning i.c.v. s.c. 60 min prior stressor to s.c. morning Mode of administration 200 mg/ kg DBA mouse DBA 200 mg/ kg 60 mg/ kg C57 mouse, high-fat diet C57 mouse, high-fat 60 mg/ kg 80 mg/ kg C57 mouse, high-fat diet C57 mouse, high-fat 80 mg/ kg 60 mg/kg C57 mouse, high-fat diet C57 mouse, high-fat 60 mg/kg 30 and 60 mg/kg rat 30 and 60 mg/kg 30 and 60 mg/kg rat 30 and 60 mg/kg 10 mg/kg rat 10 mg/kg 30 mg/kg rat 30 mg/kg 200 mg/ kg C57 mouse 200 mg/ kg 5 ng/ rat 100 ng/ rat 100 ng infusion/rat 1, 3, 5, 10, 30, and 100 ng/rat 100 ng/rat 100 ng/rat 25 mg/kg rat 25 mg/kg 25 mg/kg rat 25 mg/kg Dose Mifepristone effects on ACTH and corticosteronerodents in effects on Mifepristone antagonist Mifepristone C-125281, selective GR antagonist C-125281, selective C-108297, GR modulator Mifepristone C-118335 GR modulator/MR C-118335 GR modulator/MR C-108297, GR modulator Mifepristone Mifepristone Mifepristone Mifepristone Mifepristone Mifepristone Mifepristone Mifepristone Mifepristone Mifepristone Mifepristone 1 Table Drug

1 3 Cellular and Molecular Neurobiology

human multidrug-resistance P-glycoprotein (MDR1 Pgp) were used to examine the effect of MIF on cortisol trans- port (Karssen et al. 2001). We conclude with a discussion on the role of limbic-mesocortical circuits beyond HPA-axis regulation in prediction of treatment response to MIF.

Methods

Animals Van Haarst et al. ( 1997 ) et al. Haarst Van Ratka et al. ( 1989 ) et al. Ratka ), Block et al. ( 2018 ) et al. ( 2017 ), Block et al. Block Buckley et al. (2008) et al. Buckley Flores et al. ( 2006 ) et al. Flores Revsin et al. ( 2009 ) et al. Revsin Reference Male C57BL/6J mice, 8–10 weeks of age, were purchased from Janvier (France). Upon arrival at the animal facili- ties (Gorlaeus Laboratory, LACDR, University of Leiden, The Netherlands), mice were single housed in a tempera- ture (21 ± 1 °C) and humidity (55 ± 5%)-controlled room, with food and water ad libitum; for ten days before the start of the experiment (12–12-h light/dark cycle; lights on 0700–1900 h). During this period, mice were weighed and handled every other day. Experiments were approved by the terone duringterone both conditions corticosterone at day 7 and 14 ( n = 837, at day MIF). association of MIF with Strongest ACTH and beyond clinical efficacy cortisol cortisol ratio and ACTH/cortisol ACTH 18.00–23.00 h; Increased and cortisol 23.00–7.00 h and cortisol, slopes steeper and corticosterone levels, also in and corticosterone levels, controls Local Committee for Animal Health, Ethics and Research Increased basal ACTH and corticos - basal ACTH Increased Increased basal and stress-induced basal and stress-induced Increased Basal ACTH and cortisol elevated and cortisol elevated Basal ACTH 2 weeks post-treatment. Decreased Decreased post-treatment. 2 weeks Increased basal am and pm ACTH basal am and pm ACTH Increased Increased diabetes-induced ACTH ACTH diabetes-induced Increased Outcome of the University of Leiden. Animal care was conducted in accordance with the European Union Directive 2010/63/EU.

Study Design

The experiments were conducted with separate groups of mice. We measured (1) the 24-h circadian corticosterone secretion, following single and repeated administration of MIF (200 mg/kg; 1x/day for seven days). In addition, we collected blood samples around the time of the circadian pus 7, 14, 28, 56 i.c.v. & bilaterally dorsal hippocam - & bilaterally i.c.v. i.c.v. and s.c i.c.v. daily for 7 days, measurement on day on day measurement 7 days, for daily daily for 5 days, chronic insomnia chronic 5 days, for daily daily for 8 days for daily oral infusion, twice a day for 4 days for infusion, twice a day oral Mode of administration corticosterone peak 32 h after the last administration of MIF. (2) Exploration behavior was measured in a circular hole board. Corticosterone concentrations were analyzed before (basal) and after exposure to the circular hole. (3) Following behavioral testing, mice were decapitated and brains were prepared for measuring the expression levels of MR, GR, and CRH mRNA in the hippocampus and paraventricular nucleus of the hypothalamus (PVN).

Procedures on psychotic depression on psychotic 100 ng & 5 ng 100 ng 1200 mg/human, treatment effects 1200 mg/human, treatment 600 mg/human ( n = 5) vs control 600 mg per human, male or female 200 mg/kg mouse 200 mg/kg Dose Familiarization of mice to oat administration and drug deliv- ery procedures as described in (Dalm et al. 2008) are appli- cable to all experiments of this study.

Familiarization to Oat Administration

One week prior to the start of the experiment, a feeding cup (2.3 cm diameter x 2.5 cm high) was taped to the floor in a corner of the home cage, opposite the nest location. For familiarization, three flakes of oats (Speltvlokken, Biologis- MR antagonist MR antagonist Mifepristone Mifepristone Mifepristone 1 Table (continued) Mifepristone Drug che teelt, Graanpletterij de Halm, Netherlands; ± 140 mg)

1 3 Cellular and Molecular Neurobiology were placed in the cup on three consecutive days every other Experimental Design day, 2 h after lights on. The top of the home cage was lifted and the sawdust was removed from the cup using an air puff The circadian corticosterone secretion was determined in generated with a pipette. Next, the oats were placed into the blood samples collected via tail incision every two hour over cup using forceps to minimize human odor transfer. Thereaf- a period of 24 h. The first blood sample was taken at 1100 h, ter, the home cage was closed and the mouse was allowed to i.e., two hours after MIF or VEH was administrated, and the eat the oats undisturbed. All the oats were consumed within last at 0900 h the next day. Subsequent blood samples were 10 min. collected starting 32 h after the last administration around the circadian corticosterone peak at 1700, 1900, 2100, and Drug Delivery 2300 h. The three treatment groups (each N = 18) were divided Preparation of drug delivery via oats: One day prior to the in three subgroups each, consisting of six mice. Thus, from experiment three flakes of oats were placed in a glass vial each mouse, one blood sample was taken every six hours and the solutions containing GR antagonist or dissolvent and each time point consisted of six mice per group. During (VEH) were applied. The glass vials containing the oats the dark period, blood sampling took place under red light were kept at room temperature over night. Within 16 h, the conditions. solution was absorbed by the oats and they were dry when presented to the mice. Experiment 2: Corticosterone and Behavioral MIF (kindly provided by Corcept Therapeutics, Menlo Responses to the Circular Hole Board Park, CA, U.S.A.) was dissolved in 1 ml 0.9% NaCl con- taining 0.25% carboxymethylcellulose and 0.2% Tween 20 (VEH = dissolvent). From this solution, 50 µl was applied to Animals the oats (mice received a dose of 200 mg/kg MIF). Mice (N = 24) were randomly assigned to three treatment Hormone Assays groups (N = 8 per group): (1) single mifepristone (1xMIF); (2) mifepristone once a day on seven consecutive days The circadian corticosterone concentrations were meas- (7xMIF) or (3) VEH on seven consecutive days (VEH). ured in blood samples obtained via tail incision (Dalm et al. Oats + MIF or Oats + VEH were placed in the feeding cup 2005). Briefly, a small incision with a razor blade at the at 0900 h, and consumed within 10 min. base of the tail allowed collection of 50 µl blood within 90 s after opening of the animal’s cage. Following decapitation, Experimental Design trunk blood was collected individually in capillaries coated with potassium-EDTA (Sarstedt, Germany), stored on ice, Twenty-four hours after the last administration of MIF or and centrifuged with 13,000 rpm at 4 °C for 10 min. Plasma VEH, we took a blood sample via tail incision, and placed was stored at − 20 °C. Corticosterone concentrations were the mouse for five min on the circular hole board; the measured using commercially available radio immunoassay behavioral response was analyzed. Immediately following kits 125I-corticosterone (MP Biomedicals, Inc., NY, USA; behavioral testing, mice were decapitated. Corticosterone sensitivity 3 ng/ml). concentrations were determined in trunk blood. Brains were snap frozen in isopentane, pre-cooled on dry ice/ethanol, Experiment 1: Effect of GR Antagonism and stored at − 80 °C until further use, i.e., to determine, on Corticosterone Secretion MR, GR, and CRH mRNA expression levels in brain tissue. Thymus and adrenals were removed and weighed. Animals Circular Hole Board Mice (N = 54) were randomly assigned to three treatment groups (N = 18 per group): (1) single mifepristone (1xMIF); Apparatus A gray round plate (Plexiglass; 110 cm diameter) (2) mifepristone once a day on seven consecutive days with 12 holes (5 cm diameter, 5 cm deep) at equal distances (7xMIF); or (3) VEH on seven consecutive days (VEH). from each other, and at a distance of 10 cm from the rim Oats + MIF or Oats + VEH were placed in the feeding cup of the hole to the rim of the plate, was situated one meter at 0900 h, and consumed within 10 min. above the floor in a different experimental room than the housing room. Light conditions on the surface of the board were 120 lx. To minimize and distribute odor cues, the sur- face was cleaned with 1%HAc and the board was turned

1 3 Cellular and Molecular Neurobiology

(randomly clockwise and anticlockwise) before a mouse was The average density of six measurements for each animal tested. Behavior was recorded on videotape and analyzed was calculated. with an automated tracking system (EthoVision 3.1, Noldus Information Technology, Wageningen, The Netherlands). Statistics The position of the mouse was sampled five times per sec- ond. To calculate the distance walked, we set the minimal Data are presented as mean ± SEM. The circadian profile distance between samples to 3 cm. The following param- of corticosterone was analyzed by analysis of variance eters related to general activity, exploratory strategies, and (ANOVA) (factor: treatment) with repeated measurements, possible anxiety-related behaviors were analyzed: distance followed by Fisher’s Least Significant Difference (LSD) walked (m) on the board and in specified zones: start center post hoc test. Total corticosterone (AUC: area under the was defined as a circle of 30 cm diameter and the rim zone: a curve) over 24 h was calculated for light and dark periods ring of 4.5 cm at the outer perimeter of the plate. Parameters: of 12 h, subjected to ANOVA, with treatment and time of velocity (cm/s), number of holes visited; sequence of hole the day as fixed factors. Statistical analysis was similar as visits (serial: more than two holes in sequence; persevera- for corticosterone. Body, adrenal, and thymus weights were tion: repeatedly visiting the same hole or alternately visit- analyzed using one-way ANOVA followed by Bonferroni’s ing two neighboring holes); latency (s) to leave the center; multiple comparison post hoc test. Statistical significance latency (s) to and time spent (s) in rim zone. was accepted at p < .05.

In Situ Hybridization for MR, GR, and CRH mRNA Experiment 3: Metabolism and Membrane Transport of Mifepristone Brains were sectioned at − 20 °C in a cryostat microtome at 10 µm in the coronal plane through the level of the PVN Young adult male Wistar rats (Charles River, Germany) and dorsal hippocampus. Sections were thaw-mounted on were housed under a 12/12-h light/dark cycle with lights poly-L-lysine-coated slides (0.001%), air dried, and kept at on at 7:00 h in separate temperature (21 °C)- and humidity- − 80 °C until further use. controlled rooms. Food and drinking water were available ad In situ hybridizations using 35S-labeled ribonucleotide libitum. Before and during experiments, rats were handled probes (MR, GR, CRH) were performed as described pre- daily. Experiments were approved by the Local Committee viously (Schmidt et al. 2009). Briefly, sections were fixed in for Animal Health, Ethics and Research of the University of 4% paraformaldehyde and acetylated in 0.25% acetic anhy- Leiden. Animal care was conducted in accordance with the dride in 0.1M triethanolamine/HCl. Subsequently, brain European Union Directive 2010/63/EU. sections were dehydrated in increasing concentrations of Rats were treated with MIF for five days. Body weight ethanol. The antisense RNA probes were transcribed from was monitored throughout this period. MIF suspended in linearized plasmids containing exon-2 of mouse MR and an aqueous solution containing 0.25% carboxymethylcellu- GR, and the full length coding regions of CRH (rat). Tissue lose/0.2% Tween 20 was administered by gavage at a dose of sections (3–4 per slide) were saturated with 100 µl hybridi- 50 mg/kg once a day. Groups treated with vehicle were also zation buffer containing 20 mM Tris–HCl (pH 7.4), 50% included in all experiments. On the last day, 1.5 or 3 h after formamide, 300 mM NaCl, 1 mM EDTA (pH 8.0), 1x Den- the last injection, the animals were killed by decapitation at hardt’s, 250 µg/ml yeast transfer RNA, 250 µl/ml total RNA, the end of the day during the circadian rise of corticoster- 10 mg/ml salmon sperm DNA, 10% dextran sulfate, 100 mM one levels. Brain and plasma were collected and frozen until dithiothreitol, 0.1% SDS, 0.1% sodium thiosulfate, and sup- further use. Adrenals and thymus were also dissected and plemented with approximately 1.5 × 106 cpm 35S-labeled weighed. Plasma levels of corticosterone were determined riboprobe. Brain sections were cover-slipped and incubated using a 125I-corticosterone radioimmunoassay (MP Biomedi- overnight at 55 °C. The next day sections were rinsed in cals, Costa Mesa, CA). 2xSSC, treated with RNaseA (20 mg/ml), and washed in increasingly stringent SSC solutions at room temperature. Corticosteroid Determination in Brain and Plasma Finally, sections were washed in 0.1xSSC at 65 °C for 30 min and dehydrated through increasing concentrations Using Liquid Chromatography–Mass Spectrom- of ethanol. All age groups were assayed together. Films were etry–Mass Spectrometry (LC/MS/MS), steroid pro- opposed to Kodak Biomax MR film (Eastman Kodak Co., files were made of samples of the rat cortex and Rochester, NY) and developed. plasma. We measured levels of corticosterone, mife- Autoradiographs were digitized, and optical density of the pristone (17β-hydroxy-11β-(4-dimethylaminophenyl)- areas of interest was quantified using image analysis com- 17α-(1-propynyl)estra-4,9-dien-3-one), and its three puter software (analySIS 3.1, Soft Imaging System GmbH). main metabolites, the mono-demethylated (RU42633,

1 3 Cellular and Molecular Neurobiology

17β-hydroxy-11β-(4-monomethylaminophenyl)-17α- concentration of 15 nM. In a separate experiment, a mix of (1-propynyl)estra-4,9-dien-3-one), the didemethylated MIF and metabolites at therapeutically relevant concentra- (RU42848, 17β-hydroxy-11β-(4-aminophenyl)-17α-(1- tions was added. propynyl)estra-4,9-dien-3-one), and the hydroxylated (RU42698, 17β-hydroxy-11β-(4-dimethylaminophenyl)- In Vivo Cortisol Uptake in Brain: Effect of MIF Pretreatment 17α-(1-propynol)estra-4,9-dien-3-one) (Deraedt et al. 1985). Samples were prepared for assay by dichloromethane/etha- nol extraction essentially as previously described (Karssen To examine the in vivo effect of MIF on cortisol uptake into et al. 2001). the brain, we have treated rats orally with 100 mg/kg MIF The LC/MS/MS assays were performed on a Triple or vehicle (0.25% carboxymethylcellulose/0.2% Tween 20) (N = 7) each morning for 4 days. On the last day, all animals Stage Quadrupole mass spectrometer (Thermo Finnigan 3 TSQ Quantum, San Jose, CA, USA) with an atmospheric received a tracer dose of H-cortisol by s.c. injection 45 min pressure chemical ionization interface. A modification of after the last MIF treatment. After another 45 min, animals the method of (van der Hoeven et al. 1997) was used. The were decapitated. Trunk blood was collected and brain and analysis was performed in positive ionization mode using liver were dissected. Tissue was weighed and solubilized in selective reaction monitoring of MIF, its three main metab- Soluene-350. Together with 100 µl plasma samples, tissue olites, corticosterone, and dexamethasone. The [M + H]+ samples were counted to determine radioactivity. precursor ions were fragmented using argon as collision gas. The m/z ratios of the most abundant product ions Statistical Analysis were alternately scanned. The ion source temperature and the nebulization heater were kept at 200 °C and 400 °C, Data were evaluated by Student’s t test or ANOVA followed respectively. The voltages on the corona needle and on the by Tukey HSD post hoc test. The results of the monolayer electron multiplier were set at 10 µA. Each experiment, experiments were analyzed by Repeated Measures ANOVA. a new standard series was made in 25% methanol with Significance was taken atp < .05. concentrations ranging from 1 to 500 ng/ml of all steroids. Dexamethasone (1 µg/ml) was used as an internal stand- ard. A Surveyor LC System (Thermo Finnigan) was used Results to inject 20 µl of the standard or extraction samples. A gradient of methanol–water (containing 1 g/l acetic acid) Experiment 1: Effect of GR Antagonism changing from 50/50% to 90/10% at a flow rate of 500 µl/ on Corticosterone Secretion min separated the steroids on an ADS C­ 18 column. All samples were measured in duplo. The detection limit of Circadian Pattern of Plasma Corticosterone Level this assay was 1–5 ng/ml for each steroid. Steroid concentrations were calculated from a stand- Mice of all groups showed a circadian corticosterone rhythm ard plot of area under the curve versus concentration. The F p 2 (Fig. 1a; time (11, 165) = 35.051; < .001) as previously standard curves usually displayed an r­ of more than 0.95. described (Dalm et al. 2005). The corticosterone secretion Presented data are corrected for recovery of dexamethasone, of control mice increased from 1500 h onwards, with peak which was in the order of 25–50%. levels (± 100 ng/ml) at the end of the light phase and the beginning of the dark phase (between 1700 and 2100 h). Transepithelial Transport and Inhibition Studies Interestingly, the frequency of MIF administration affected the course of the circadian rhythm (time*group: F (22, In order to examine the inhibitory action of MIF on Pgp- 165) = 15.992; p < .001). Corticosterone concentrations mediated cortisol transport, we used monolayers of the por- in 1xMIF-mice were significantly higher from 1100 until cine kidney epithelial cell-line LLC-PK1, and LLC-PK1 0100 h (p < .01), reaching and maintaining peak levels from cells stably transfected with cDNA of the human MDR1 1300 until 2300 h (± 300 ng/ml). Around 2300 h, concentra- gene encoding P-glycoprotein (LLC-PK1:MDR1) as pre- tions readily declined until there was no difference in corti- viously described (Karssen et al. 2001). Cells obtained costerone concentration at 0300 h versus control and 7xMIF- from the American Type Culture Collection (Manassas, administrated mice. There was a sudden significant increase VA) were kindly provided by the Dutch Cancer Institute versus controls (p = .001) and 7xMIF mice (p = .013), at (Amsterdam, The Netherlands) (Schinkel et al. 1995). MIF 0500 h. In contrast, repeated MIF administration did not was added at a final concentration of 10 or 100 µM one 3 boost the concentrations of corticosterone as was observed hour before the addition of H-cortisol (Amersham Phar- for 1xMIF-administrated mice; the time course was similar macia Biotech, UK; specific activity 63 Ci/mmol) at a final

1 3 Cellular and Molecular Neurobiology

Fig. 1 a Circadian secretion of corticosterone in ng/ml measured secretion in ng/ml during the light and dark period of the day, deter- every 2 h in blood plasma of male mice C57BL/6J that received mined as area under the curve (AUC); ng/ml. Data are presented as RU38486 (MIF) once (1xMIF) or for seven days (7xMIF). Mice were mean ± SEM; p < .05 * versus other groups, # within groups, ~ 7xMIF entrained in a 12–12-h light–dark cycle (dark phase from 1900 to versus VEH 0700 h represented by the gray-shaded area). b Total corticosterone

to VEH mice. Overall, there was a main effect of treatment due to the high corticosterone concentrations in the 1xMIF mice (F (2, 15) = 550.923; p < .001).

Total amount corticosterone

The total amount of corticosterone calculated as area under the curve (AUC) over 24 h showed a main effect of treat- ment (Fig. 1b AUC: F (2, 17) = 392.094; p < .001). AUC corticosterone during the dark period (1900–0700 h) was higher than during the light period (0700–1900 h) in VEH Fig. 2 Corticosterone (ng/ml) secretion during the circadian peak and 7xMIF mice (paired t-test; both p < .01). 1xMIF mice in mice, 32 h after last administration of RU38486 (MIF), 1xMIF, had similar high AUC corticosterone levels during the light 7xMIF, or VEH (dark phase from 1900 to 2300 h represented by the and dark periods, both significantly higher than VEH and gray-shaded area). Data are presented as mean ± SEM 7xMIF mice. Interestingly, AUC corticosterone was low- est during the light period of 7xMIF mice (p < .039 versus secreted less corticosterone than VEH (p = .007) and 7xMIF VEH) due to the low corticosterone concentrations measured mice (p = .008). No statistical difference was found for cor- from 1500 till 1700 h. ticosterone secretion patterns of VEH and 7xMIF groups.

Experiment 2: Corticosterone and Behavioral Corticosterone Around the Circadian Peak: 32 h After Responses to the Circular Hole Board Mifepristone Administration Basal and Novelty‑Induced Corticosterone Secretion Treatment effects were found around the time of the circa- dian peak (Fig. 2, 1700–2300h; F (2, 15) = 6.308; p = .01). Basal resting as well as novelty-induced corticosterone were Thirty-two hours after the last administration, 1xMIF mice affected 24 h after the last treatment (Fig. 3; treatment F (2,

1 3 Cellular and Molecular Neurobiology

Expression of MR, GR, CRH mRNA in Hippocampus and PVN

Hippocampal MR mRNA expression was differentially affected by treatment, 24 h post-administration, across all subfields (Fig. 4; treatment – DG: F (2, 23) = 11.005; p = .001; CA1: F (2, 23) = 12.887; p = .001; CA2: F (2, 23) = 14.267, p = .001; CA3: F (2, 23) = 11.550; p = .001). MR mRNA expression was reduced across all subfields in 1xMIF-mice compared to VEH and 7xMIF-mice (p < .05). Repeated MIF administration increased MR mRNA expres- sion in the CA2 specifically versus VEH and 1xMIF-mice (p = .016 and p = .001, respectively). Neither GR nor CRH mRNA expression in hippocampus and PVN were affected by treatment (data not shown).

Exploration on the Circular Hole Board Fig. 3 Basal and novelty (5 min exposure to the circular hole board)- induced corticosterone (ng/ml) were determined in mice, 24 h after last administration of VEH, 1xMIF, or 7xMIF. Data are presented as Twenty-four hours after administration, the behavioral p # mean ± SEM; < .05 * versus other groups, within groups response differed during five min exploration on the cir- cular hole board (Table 2: MANOVA: F (20, 26) = 3.772; 44) = 17.175; p < .0001; time F (1, 44) = 45.980; p < .0001; p = .001). Following initial slower movement out of the treatment*time F (2, 44) = 17.626; p < .0001). Basal rest- central start position, 1xMIF mice showed hyperactivity: ing corticosterone differed significantly between the groups they walked longer distances, with a faster speed of mov- (F (2, 23) = 14.656; p < .001) and was lower in both MIF- ing, visited more holes, and made more rim dips (vs. VEH treated groups than in VEH mice (p < .001). Basal corticos- and 7xMIF-mice: p < .05). Interestingly, 7xMIF-mice made terone of 1xMIF and 7xMIF mice was comparable. After more use of a serial search strategy (vs. VEH mice: p = .05). five min on the circular hole board, corticosterone was increased in all groups compared to baseline, however to a different degree (F (2, 23) = 19.074; p < .0001). Corticos- Other Physiological Measures terone levels in 1xMIF were 300% of the VEH group and 700% of the 7xMIF group (both p < .0001); corticosterone Treatment did not influence body weight. Adrenal weight (F of the VEH group was about twice as high as in the 7xMIF (3, 34) = 3.733; p = .035) was highest in both MIF groups, group (p < .05). but significantly higher in 7xMIF than in VEH (p = .005): adrenals in mg, mean ± SEM: VEH 23.5 ± 2.8; 1xMIF

Fig. 4 Expression of MR mRNA, measured as optical density (OD) in the hippocam- pal subfields dentate gyrus (DG), CA1, CA2, and CA3, 24 h after last administration of VEH, 1xMIF, or 7xMIF. Data are presented as mean ± SEM; p < .05 * versus other groups, # within groups

1 3 Cellular and Molecular Neurobiology

Table 2 The behavioral response during five min circular hole board Experiment 3: Metabolism and Membrane Transport exposure, 24 h after the last administration with RU38486 (MIF) of Mifepristone VEH 1xMIF 7xMIF In Vivo Steroid Uptake in Brain General activity Distance walked (m) 7.9 ± 0.7 15.0 ± 1.9* 7.5 ± 0.9 In experiment 1, the steroid profiles of rat brain and plasma Speed of moving (cm/s) 8.6 ± 0.4 11.6 ± 0.8* 9.9 ± 0.3 made with LC/MS/MS showed that after oral treatment Total hole visits 14.8 ± 2.1 24.8 ± 1.8* 17.0 ± 2.4 with 50 mg/kg/day, MIF could not be detected in plasma Search strategy # at either 1.5 or 3 h after the fifth and last treatment. Neither Latency (s) from center 8.4 ± 1.4 14.0 ± 1.4 11.8 ± 2.1 # could any of its metabolites. In contrast, MIF was detect- Latency (s) first hole visit 13.9 ± 0.8 16.0 ± 2.6 18.4 ± 0.7 # able in the brain with no significant difference between % Serial 16.5 ± 5.2 28.2 ± 3.6 36.6 ± 10.2 both time points, although individual variability in brain % Perseveration 48.6 ± 5.8 39.3 ± 5.2 52.1 ± 5.1 MIF levels was quite high. Comparable but more consist- Anxiety related ent levels were found for RU42633 in brain after treatment Latency (s) to rim 63.0 ± 13.1 55.1 ± 12.4 69.9 ± 8.5 with MIF (Fig. 5). The mono-demethylated metabolite was Number of rim dips 12.8 ± 1.4 18.4 ± 2.0* 11.1 ± 1.2 also present at very low, but detectable levels in plasma. Number of boli 1.1 ± 0.7 0.8 ± 0.4 1.3 ± 0.8 Moreover, plasma levels significantly correlated with brain 2 Data are presented as mean ± S.E.M.; p < .05 * versus other groups; # levels for MIF-treated animals (r = 0.52, p < .01). Corticos- versus VEH terone levels in brain also correlated strongly with plasma Bold italic indicates significant differences levels (r2 = 0.48, p < .01) as determined with LC/MS/MS, and the latter values were validated with 125I-corticosterone radioimmunoassay (r2 = 0.89, p < .01; data not shown). After 31.3 ± 4.3; 7xMIF 39.3 ± 2.9. Thymus weight was lower in oral treatment with MIF, corticosterone levels increased sig- both MIF groups, but passed statistical significance (F (3, nificantly in plasma (F (2,15) = 7.94, p < .01) (Fig. 5). In 34) = 3.100; p = .059): thymus in mg, mean ± SEM: VEH brain, the increase did not reach statistical significance. MIF 411.0 ± 38.9; 1xMIF 371.5 ± 15.1; 7xMIF 321.1 ± 23.6. treatment did not affect adrenal or thymus weight nor body weight.

Fig. 5 Steroid levels at 1.5 or 3 h after the last oral administration of of rats treated with MIF (F(2,15) = 7.94, p < .01) compared to vehicle 50 mg/kg mifepristone; levels of mifepristone were undetectable in -treated rats. The concentrations of RU42848 and RU42698 were plasma, but clearly detectable in brain although with high variability. below the detection limit in both plasma and brain. N = 4–6, shown Brain RU42633 levels were significantly higher in MIF-treated ani- is mean + sem, * p < .05, Tukey post hoc test. Note the difference in mals compared to vehicle treated rats (F(2,15) = 13.12, p < .01). Corti- scale costerone levels were significantly higher in plasma but not in brain

1 3 Cellular and Molecular Neurobiology

Fig. 6 a Fraction of activity of 3H-cortisol present in medium at dif- port of 3H-cortisol in monolayers of MDR1-transfected cells is inhib- ferent time points after adding 15 nM 3H-cortisol to the opposite ited and not different from transport of cortisol in monolayers of hosts compartment at t = 0 in absence or presence of MIF. Transepithelial cells. Data are presented as mean ± sem of three wells. MIF did not transport from basal to apical compartment and vice versa was meas- affect cortisol transport in untransfected monolayers (data not shown). ured in MDR1-transfected LLC-PK1 monolayers. Repeated measures b A mix of MIF and its three main metabolites at therapeutically rel- ANOVA showed a significant time * cell type * MIF * direction of evant concentrations(see text) inhibits the transport of 3H-cortisol in transport interaction (p < .01). In the presence of 10 µM MIF, trans- MDR1-transfected monolayers

Inhibition of Cortisol Transport In Vitro MIF was 0.08 versus 0.11 nCi/mg tissue. Although the amount of radioactivity in MIF-treated animals was not The in vitro experiments confirmed our previous observa- significantly enhanced in any tissue or plasma, the blood/ tions (Karssen et al. 2001) that in MDR1-monolayers, cor- brain ratio of radioactivity administered as 3H-cortisol was tisol was transported in a highly polarized fashion (Fig. 6a). significantly increased in MIF-treated animals, indicat- In presence of 10 µM MIF, this transport was inhibited. ing a facilitation of cortisol uptake into the brain. The fold ANOVA followed by post hoc analysis shows that at t = 4 increase (1.9x) is less than in Pgp knock-out mice as previ- in presence of MIF, MDR1-transfected monolayers were ously determined (3.5x) (Karssen et al. 2001). not different from untransfected monolayers with regard to transport of 3H-cortisol, while both were statistically differ- ent from the untreated MDR1 monolayers. 100 µM MIF was Discussion not able to further enhance the inhibitory action on cortisol transport (data not shown). The current data extend in male mice and rats the notion that As the three main metabolites of MIF are structurally a single challenge with MIF interferes with glucocorticoid closely related to MIF, they may inhibit Pgp as well. There- feedback causing a long-lasting elevation of corticosterone fore, we tested the ability to inhibit cortisol transport of secretion (Ratka et al. 1989; Dalm et al. 2008). Moreover, we a mix of MIF and metabolites at therapeutically relevant found that at 24 h after ingestion of MIF, stress responsivity concentrations (Lähteenmäki et al. 1987). A mix of MIF, is profoundly enhanced. However, upon repeated high-dose RU42848, RU42698 (2.5 µM each), and 6 µM RU42633 MIF (200 mg/kg) daily administration, the circadian- and affected cortisol transport in MDR1 monolayers to a similar stress-induced HPA-axis activity are abolished. This appar- extent as 10 µM MIF alone (Fig. 6b), whereas 2.5 µM MIF ent GR agonism of MIF upon repeated administration was alone inhibited transport to a minor extent only (data not previously also reported (Havel et al. 1996; Wulsin et al. shown). 2010; van der Veen et al. 2013; Solomon et al. 2014; van den Heuvel et al. 2016; Nguyen et al. 2017, 2018; Kroon et al. 2018); however, with the lower doses in these studies, 3 In Vivo Uptake of H‑Cortisol: Effect of Mifepristone stress responsivity was attenuated rather than abolished (see Pretreatment also Fig. 5). At the same time, the weight of the adrenals was increased in our study. Adrenal weight was not altered with In a preliminary experiment, at 45 min after 3H-cortisol, the lower 20-fold lower dose of MIF over 5 days (Wulsin blood plasma radioactivity was Veh (n = 7) versus MIF et al. 2010) or fourfold lower dose (exp 3), and decreased (n = 8), 0.88 versus 0.65 nCi/ml, while in brain, Veh versus if a 3–10-fold lower dose of MIF was administered over a

1 3 Cellular and Molecular Neurobiology longer time period of two weeks (Havel et al. 1996; Kroon et al. 2018). Thymus weights were consistently decreased, although not significant in the current study. As mentioned, this daily administration showing agonism contrasts with the continuous i.c.v. and twice a day high dose of MIF in that the latter two conditions produce in rodents disinhibition of the HPA axis in the circadian rise and following stress (van Haarst et al. 1996; Revsin et al. 2009). These findings raise the question how the disinhibitory effect of acute MIF can abruptly change into an enduring inhibitory one upon repeated daily administration, particu- larly at doses in the range of therapeutic efficacy. A second, equally important question is, whether the current data con- tribute to understanding the apparent reset of the stress sys- tem achieved with MIF (Ratka et al. 1988; Hu et al. 2012). Various factors may contribute to this switch of GR antago- Fig. 7 Hypothesized MIF-induced facilitation of cortisol brain uptake nism to an apparent agonism, and ability to reset. These through inhibition of the efflux transporter P-glycoprotein at the include (i) MIF kinetics: metabolism and penetration to GR blood–brain barrier. Under normal conditions, cortisol is hampered to enter the brain due to active outwards directed transport at the blood– feedback sites in brain and pituitary, (ii) dynamics of HPA- brain barrier mediated by P-glycoprotein (Karssen et al. 2001). In the axis feedback regulation, (iii) differential signaling routes of presence of MIF, this efflux is blocked facilitating the uptake of cor- MIF- and cortisol-occupied GR beyond HPA-axis regula- tisol into the brain. The ensuing increased cortisol concentration will tion, and (iv) the brain MRs. not lead to increased activation of GR, since this receptor is blocked by the high concentrations of MIF. However, increased activation of the MR is predicted to affect cognitive performance and neuroendo- MIF Kinetics crine regulation. X indicates blockade of Pgp by MIF, which facili- tates cortisol penetration through the blood–brain barrier and block- MIF is rapidly cleared because in rodents the antagonist is ade of cortisol binding to GR. Dotted line is preferred cortisol route after Pgp blockade by MIF not bound to plasma proteins and rapidly metabolized, while in human blood MIF is extensively protected to degrada- ɑ tion because of binding with high affinity to 1 glycoprotein In AtT 20 cells, MIF translocated the GR to the cell (Heikinheimo et al. 1987). We found that after a 50 mg/kg nucleus and induced DNA binding and MIF potentiated the rat dose orally given during 5 days, after the fifth and last cell nuclear translocation and DNA binding of corticoster- administration MIF is already depleted from the circulation one (Spiga et al. 2011). in vivo, MIF (20 mg/kg)-admin- in a 90-min post-ingestion interval, while low amounts of istered s.c. to ADX rats induced translocation and DNA the antagonist and its metabolite RU42633 are retained in binding of GR to nuclei of hippocampus 60 min following the brain for at least 3 h and in the same concentration range administration (van Eekelen et al. 1987; Spiga et al. 2011). as corticosterone. The retention of the MIF–GR complex in pituitary nuclei The in vitro experiments using monolayers of pig kidney was even more than twofold higher than in hippocampus. epithelial cells (LLC-PK1) stably expressing human MDR1 Pretreatment of the ADX rats with MIF did not inhibit but Pgp (Karssen et al. 2001) showed that MIF was capable of rather potentiated the nuclear retention of corticosterone facilitating the transport of cortisol. We have previously (3 mg/kg) in hippocampal and pituitary nuclei; however, shown that the uptake of dexamethasone and cortisol is the effect was near trend level (p < 0,06) (Spiga et al. 2011). hampered by Pgp at the blood–brain barrier in rodents and Whether this also occurs in vivo in intact animals with cir- humans (Meijer et al. 1998; Karssen et al. 2001; Mason et al. culating corticosterone is not known. 2012). MIF blocks Pgp, so cortisol cannot be bound and thus In conclusion, the data show that in rodents MIF is rap- is not exported by the mdr transporter, hence a higher uptake idly cleared from the circulation but retained and bound to and retention of cortisol. In an in vivo experiment, it was GR in hippocampal and pituitary cell nuclei and DNA. Most indeed demonstrated that a higher uptake of radioactivity interestingly, MIF is a substrate for mdr transporters in the 3 in brain relative to blood occurred if H-cortisol is infused blood–brain barrier, and therefore in rodents the uptake and in animals pretreated with a high dose of MIF. The finding retention of exogenous cortisol is facilitated (Gruol et al. shows that repeated MIF could increase accumulation of the 1994; Lecureur et al. 1994; Gaillard et al. 1984; Karssen endogenous glucocorticoids in the brain of adrenally intact et al. 2001, 2005, 2004, 2002). Collectively, these data sup- animals by blocking Pgp at the blood–brain barrier (Fig. 7). port the idea that at least in rodents very high doses of MIF

1 3 Cellular and Molecular Neurobiology are needed at the pituitary and brain level to overcome its while the transrepression-dependent glucocorticoid negative rapid metabolism and the facilitated uptake and retention feedback on ACTH release is mostly intact (Reichardt et al. of cortisol. 1998; Oitzl et al. 2001). Also, partial agonism of MIF via classical transactivation cannot be excluded in some cell HPA‑Axis Dynamics types, and over time. For instance, in simple reporter gene assays, the extent of partial agonism could be doubled by One possibility to explain the sudden refractoriness to overexpression of the SRC-1A coactivator (Meijer et al. repeated MIF administration is that the pituitary ACTH 2005) that is relatively abundant in the core of the HPA axis stores are depleted. Such depletion is a common observa- (Meijer et al. 2000). Prolonged exposure to either continu- tion following ADX when after the initial surge in readily ous MIF, or alternating MIF and corticosterone in the mice releasable ACTH minimum pituitary levels of irACTH are studies, therefore has an unpredictable outcome, that may attained after 1–3 days. Meanwhile, synthesis of POMC is moreover develop over time, given that the expression of stepped up and a condition of several folds higher ACTH coactivators themselves may be regulated over time (Meijer level is reached 1 week after ADX, which still can be further et al. 2005). elevated by stress exposure (van Dijk et al. 1981; Jacobson et al. 1989). However, ACTH levels were increased dur- ing twice a day of MIF for 4 days (Revsin et al. 2009). It Mineralocorticoid Receptors is therefore unlikely that ACTH exhaustion has occurred, but for a conclusive answer, however, the effect of MIF on Irrespective of phasic or continuous GR blockade, corticos- pituitary and plasma ACTH needs to be studied still on a terone binding to the MR is not hampered by MIF (Reul day-by-day basis. et al. 1990). MR is known to mediate control over appraisal An alternative explanation can be a recurrent negative processes, behavioral reactivity to novel experiences, and feedback. After a single MIF challenge, corticosterone the onset of HPA-axis activity (Joëls et al. 2008; de Kloet remained elevated for 16 h, while the genomic effects will and Joëls 2017; Joëls and de Kloet 2017). The changes in persist even longer. Indeed, at 24 h, the 1xMIF-treated ani- corticosterone action via MR activation were reflected by the mals were capable of mounting a profound corticosterone lower expression of hippocampal MR mRNA for the 1xMIF- response. Our study suggests a recurrent pattern of GR- group in all subregions, whereas expression was increased in mediated actions including negative feedback, which are the CA2 region of the hippocampus for 7xMIF-group. Inter- transiently interrupted by daily application of the GR antag- estingly, GR blockade after MIF treatment per os for three onist. We propose therefore that the HPA axis progressively weeks, increased total hippocampal MR mRNA expression adapts to this daily cycle of GR blockade and subsequent by 1.5 compared to controls (Bachmann et al. 2003). There- GR activation. Hence, during the seventh day of GR antago- fore, it would be of interest to determine longitudinal effects nist administration, the circadian corticosterone pattern may of repeated GR blockade on MR function, particularly since have become similar to that observed in control mice. Natu- previous studies have clearly shown that MR and GR interact rally, this assumption needs proof by measuring the effect of in control of HPA-axis activity (Spencer et al. 1998). MIF on basal and stress induced on a day-by-day basis, and Four days of MIF with the last administration given to compare this study with twice a day or continuous MIF 90 min prior to the initial test promoted an active coping administration thought to maintain disinhibition. style (Wulsin et al. 2010). Acute MIF interfered with reten- tion of this acquired immobility response, suggesting mem- MIF and Cortisol Signaling Via GR ory impairment (de Kloet et al. 1988). Chronic MIF, leav- ing MR active, improved memory performance (Oitzl et al. The apparent agonism of MIF that is observed in some con- 1998). In line with this phenotype of a relative brain MR texts may differ from the endogenous agonism via cortisol excess, studies with mouse mutants overexpressing MR in in several ways. First, it has been proposed that MIF–GR is limbic forebrain revealed enhancement of memory (Fergu- much more potent in exerting transcriptional transrepression son and Sapolsky 2007; Lai et al. 2007; Kolber et al. 2008), than transactivation activity (Heck et al. 1994). This effec- perseveration of learned behavior (Harris et al. 2013), and tively could mean that—once endogenous corticosterone reduction of anxiety (Rozeboom et al. 2007). levels are low—part of the GR-dependent negative feedback Previously, the pharmacological blockade of MR had does get activated by MIF via transrepression (De Bosscher also shown altered appraisal processes and selection of the et al. 2003, 2016). This is consistent with the HPA-axis phe- appropriate behavioral response, i.e., search strategy (Oitzl notype of the GR­ dim/dim mouse, where transcription of the and de Kloet 1992; Oitzl et al. 1994, 1997; Schwabe et al. Pomc gene and memory performance are disrupted, because 2010, 2013). In the current study, twenty-four-hours follow- of its inability for GR homodimerization and transactivation, ing repeated GR antagonism, mice used the serial search

1 3 Cellular and Molecular Neurobiology strategy more often, compared to controls. This strategy The current data raise the possibility that MIF’s pri- increases the likelihood that the animals will visit all possi- mary target is extrahypothalamic. A recent study demon- ble escape routes that the circular hole board provides (dur- strated in rats that chronic MIF released from a 150 mg s.c. ing the exploration trial, all holes were closed). Indeed, the implanted pellet reduces (compulsatory) alcohol intake in choice of applied strategy does affect performance in spa- dependent rats (Vendruscolo et al. 2012). These effects tial learning trials (Dalm et al. 2000). In the current study, were reproduced by i.p. administration of 30–60 mg MIF following 1xMIF, mice were initially slower to move away at 90 min prior to alcohol intake as well as after 10–30 µg from the start center, but subsequently hyperactive on the local administration bilaterally in the central amygdala. The circular hole board. This could indicate a change in the level data were reproduced in humans with MIF and the selective of anxiety induced by previous GR antagonism. If so, then GR antagonist C-113176 (Vendruscolo et al. 2015). It was the effect is transient as repeated GR antagonism did not also found in rats that alcohol withdrawal and protracted induce any of the above described features. abstinence produced changes in GR mRNA and bio-active The elegant study by (Wulsin et al. 2010) revealed that phosphorylated GR rather than MR mRNA expression in the one week treatment with a twentyfold lower dose of MIF i.p. limbic-prefrontocortical and mesolimbic dopaminergic sys- (10 mg/kg rat) produced an attenuated HPA-axis response tem. Using a systems biology approach, GR was identified as to a forced swim stressor. Interestingly, this course of MIF a master controller of downstream gene regulatory networks treatment also evoked a differential pattern of activation and in these extrahypothalamic regions (Repunte-Canonigo et al. inhibition of central inputs to the PVN. The ventral subicu- 2015). lum of the hippocampus and all regions of the medial frontal When it concerns diseases precipitated by psychological cortex showed enhanced stress-induced c-Fos activity after stress, the ability of MIF and related compounds to read- daily GR blockade, while the c-Fos response was reduced in ily re-establish the setpoint of the stress system is of prime other subregions of the hippocampus and in the amygdala. importance. Such a reset would require restoration of an These data suggest that MIF enhanced inhibitory and sup- imbalance in MR/GR-mediated processes in the brain. A pressed excitatory inputs to the PVN that collectively may striking example is found in rats showing a deterioration of contribute to the downregulation of HPA-axis activity. To cellular and behavioral phenotype during chronic unpredict- exert these effects, MIF likely may have recruited distinct able stress, which is entirely restored with a single course of cocktails of co-regulators of the GR in the corticosteroid MIF (Hu et al. 2012). Transcriptome analysis (Datson et al. target neurons of the limbic brain (Meijer et al. 2000; Mei- 2012) revealed that in laser-dissected hippocampal dentate jer et al. 2005; Lachize et al. 2009; Ronacher et al. 2009; gyrus of these chronically stressed animals, MIF treatment Zalachoras et al. 2013). In conclusion, as a result of GR affected 107 genes, which were mostly different from the blockade, the MR becomes relatively more activated. This ones observed in the stressed group. We found that CREB- combination of increased MR activity and blockage of GR binding protein (CREB-BP) was normalized by MIF treat- may result in an altered function of limbic-frontocortical ment to levels observed in control animals. Next to the GR, afferents and reset of neuroendocrine control of the HPA CREB-signaling, therefore, may play a central role in medi- axis. ating the chronic stress effects on neurogenesis, LTP, and calcium currents in the dentate gyrus (Karst and Joëls 2003; Implications for Clinical Studies van Gemert and Joëls 2006; Datson et al. 2012, 2013) and perhaps in other (meso)limbic-frontocortical regions as well. One obvious question is how the current findings in the In humans, bio-availability of MIF is extended because rodent may contribute to the clinical observations show- the steroid circulates bound to ɑ1 glycoprotein and therefore ing high-dose MIF efficacy. Recent studies referred to the has a much longer half-life than in rodents, where this bind- need of circulating MIF levels exceeding 1637 ng/ml blood ing protein is absent (Heikinheimo et al. 1987). Moreover, in with a dosage of 1200 mg/day for 7 days in order to achieve recent studies it was noted that although there is an obvious the strongest association with treatment response in psy- association with MIF’s action on the HPA axis, it cannot be chotic depression. This association in treatment response excluded that the compound’s pharmacology is due to block- was followed by a weaker, but still significant association ade of GRs in limbic-frontocortical and mesocortical dopa- with increased basal ACTH and cortisol levels (Block et al. minergic circuits involved in mood, anxiety, reward, and 2018), which lasted several weeks beyond the termination of motivation. This is reinforced by the MIF-induced change in MIF treatment. Before the conclusions on reset of the stress stress circuit activation favoring inhibitory inputs to the PVN system from animal experiments can be extended, more (Wulsin et al. 2010), the systems biology approach pointing data on stress responsivity in clinical studies are required, to the GR in these regions as master controller (Repunte- however. Canonigo et al. 2015) and the MR:GR balance hypothesis (de Kloet et al. 2018). Although the required circulating

1 3 Cellular and Molecular Neurobiology

MIF levels beyond 1637 ng MIF/ml blood certainly are an interesting new ligand is C-118335, a mixed GR modula- important first step, more data on HPA-axis regulation and tor/MR antagonist (100 mg/ kg male rat), which displayed extrahypothalamic effects as well as the role of (epi)genetic clearly GR agonistic activity in gene regulation and HPA- factors are needed to predict treatment response in psychotic axis suppression after a single s.c. injection, but blocked depression. memory storage (Atucha et al. 2015), while being inactive on coping style (Nguyen et al. 2017, 2018). Hence, more Perspectives selective GR modulators will be important for targeted treatment. MIF (or Korlym®) is indicated for treatment of hypergly- cemia during Cushing’s Syndrome and was found to cause a clinically significant metabolic improvement with less Concluding remarks depressive symptoms, improved cognitive performance, and an increased quality of life (Fleseriu et al. 2012). This indi- The reset of stress system activity by MIF could be due to cation is based on the correction of glucocorticoid-induced the following factors: (1) The detrimental effects of high cor- glucose intolerance and diabetes mellitus (Beaudry et al. ticosteroid concentrations via GR activation are prevented 2014; Teich et al. 2016; van den Heuvel et al. 2016; Moraitis by GR antagonism. This approach has obvious benefit dur- et al. 2017; Kroon et al. 2018; Meijer et al. 2018). ing continuous or high-frequency blockade of the GR. (2) Glucocorticoid action is of course needed for allocation The high circulating corticosterone concentrations may of energy substrates to tissues in need during coordination have caused a long-lasting rebound suppression of HPA- of circadian events, immune and inflammatory defense axis activity, particularly since MIF is rapidly cleared. (3) reactions, stress coping, and adaptation (Picard et al. 2014; MIF may become an agonist in transrepression (Heck et al. Hollis et al. 2015). If defense reactions essential for health 1994) and promote cell- and context-dependent recruitment exceed control by endogenous corticosterone and cortisol, of coactivators and co-repressors in a cell-specific fashion they may become damaging themselves (Munck et al. 1984; in brain stress circuitry (Meijer et al. 2018; Zalachoras et al. Sapolsky et al. 2000). The exogenous (synthetic) glucocor- 2013, 2016). (4) Enhanced brain MR activation may become ticoids are very effective anti-inflammatory and immune prominent relative to the MIF-modulated GR. The altered suppressive agents. However, the adverse effects on energy MR:GR balance could be part of a compensatory mecha- metabolism, pituitary ACTH release, and brain function nism producing altered patterns of inhibitory and excitatory may become a serious concern. Important progress has been circuits underlying reset of stress system activity (De Kloet made by exploiting SGRM’s for identification of tissue and et al. 1998; de Kloet et al. 2005, 2018; Sousa 2016; Wulsin context-specific modulation of GR-mediated processes. This et al. 2010). This includes the role of MR and GR as homo- is important because the ‘golden bullet’ (Meijer et al. 2018) or heterodimers in binding to GRE’s, transcription factors, would be a compound that solely treats inflammatory and and co-regulators (Mifsud and Reul 2016; Van Weert et al. immune processes without side effects, or that solely restores 2017; Matosin et al. 2018; Meijer et al. 2018) as well as their deregulated metabolism, or that targets only central circuits rapid non-genomic actions (Di et al. 2003; Karst et al. 2005, underlying depression. 2010; Groeneweg et al. 2012; Hill and Tasker 2012). Potential new applications of the SGRM’s are being Chronic stress models of depression display glucocorti- tested. For instance, C-108297 (3,5 mg/mouse) for 4 days coid resistance caused by imbalance in the afferent limbic- given to Wobbler mice, an animal model of human amyo- prefrontocortical pathways innervating the hypothalamic trophic lateral sclerosis (ALS), normalizes indices for neu- PVN and mesolimbic dopaminergic system (Ulrich-Lai and rogenesis and facilitates recovery from a pro-inflammatory Herman 2009). The striking finding with MIF and related phenotype in hippocampus (Meyer et al. 2014). The selective compounds is that the high doses required to affect pituitary GR antagonist C-113176 displayed a similar protective phar- and brain regulation of the HPA axis seem capable of read- macology in the Wobbler mice, and additionally prevented ily resetting stress system activity. It is a great challenge to spinal cord pathology, while displaying anti-inflammatory discover if reset of stress system activity is the cause or the and anti-glutamatergic activity (Meyer et al. 2018). consequence of MIF’s efficacy in the treatment of stress- MIF (30 mg/kg rat ip), and the much more effective related neuropsychiatric disorders. C-108297 (20 mg/kg) and C-113176 (10 mg/kg) at a dose of 20 and 10 mg/kg, respectively, twice a day for 5 days, Acknowledgements The inspiring leadership of Melly S. Oitzl in the research on stress, brain, and behavior is gratefully acknowledged. all reversed the precipitation of Alzheimer pathology and Technical assistance of Ewald Engst and Julia Straker is gratefully cognitive impairment in an animal model generated by hip- acknowledged. pocampal amyloid-β25−35 administration, while normalizing plasma corticosterone levels (Pineau et al. 2016). Another

1 3 Cellular and Molecular Neurobiology

Author Contributions SD and AMK performed the experiments as part Blasey CM, Block TS, Belanoff JK, Roe RL (2011) Efficacy and of their PhD research, wrote the methods and results section, and made safety of mifepristone for the treatment of psychotic depression. the figures. OCM evaluated the manuscript for mechanistic underpin- J Clin Psychopharmacol 31:436–440. https​://doi.org/10.1097/ ning of the GR antagonist action and wrote part of the discussion. JKB JCP.0b013​e3182​23919​1 incorporated text on specific knowledge of mifepristone and novel GR Block T, Petrides G, Kushner H et al (2017) Mifepristone plasma level modulators in the text. ERdK designed the study, wrote the overview and glucocorticoid receptor antagonism associated with response on GR antagonists in the introduction, and integrated the results in the in patients with psychotic depression. J Clin Psychopharmacol abstract and discussion sections. All authors critically reviewed the 37:505–511. https​://doi.org/10.1097/JCP.00000​00000​00074​4 manuscript and gave approval to the final version of the manuscript. Block TS, Kushner H, Kalin N et al (2018) Combined analysis of mifepristone for psychotic depression: plasma levels associated Funding Funding was provided by NWO #015.01.076 (SD) and the with clinical response. Biol Psychiatr 84:46–54. https​://doi. Royal Netherlands Academy of Arts and Sciences (ERdK). Corcept org/10.1016/j.biops​ych.2018.01.008 Therapeutics has provided Mifepristone for the studies. Dalm S, Grootendorst J, R de KE SOM (2000) Quantification of swim patterns in the Morris water maze. Behav Res Methods Instrum Comput. 32:134–139 Compliance with Ethical Standards Dalm S, Enthoven L, Meijer OC et al (2005) Age-related changes in hypothalamic-pituitary-adrenal axis activity of male Conflict of interest SD and AMK report no conflict of interest. ERdK C57BL/6J mice. Neuroendocrinology 81:372–380. https​://doi. is on the Scientific Advisory Board of the DynaCorts Group and owns org/10.1159/00008​9555 stock of Corcept Therapeutics. OCM receives funding from Corcept Dalm S, Brinks V, van der Mark MH et al (2008) Non-invasive Therapeutics. JKB is an employee of Corcept Therapeutics, which de- stress-free application of glucocorticoid ligands in mice. J velops novel GR ligands. Neurosci Methods 170:77–84. https​://doi.org/10.1016/j.jneum​ eth.2007.12.021 Open Access This article is distributed under the terms of the Crea- Datson NA, Speksnijder N, Mayer JL et al (2012) The transcriptional tive Commons Attribution 4.0 International License (http://creat​iveco​ response to chronic stress and glucocorticoid receptor blockade mmons.org/licen​ ses/by/4.0/​ ), which permits unrestricted use, distribu- in the hippocampal dentate gyrus. Hippocampus 22:359–371. tion, and reproduction in any medium, provided you give appropriate https​://doi.org/10.1002/hipo.20905​ credit to the original author(s) and the source, provide a link to the Datson NA, van den Oever JME, Korobko OB et al (2013) Prior Creative Commons license, and indicate if changes were made. history of chronic stress changes the transcriptional response to glucocorticoid challenge in the dentate gyrus region of the male rat hippocampus. Endocrinology 154:3261--3272 de Kloet ER (1991) Brain corticosteroid receptor balance and home- ostatic control. Front Neuroendocrinol 12:95–164. https://doi.​ References org/10.1080/09614​52070​14696​17 de Kloet ER, Joëls M (2017) Brain mineralocorticoid receptor Arp JM, ter Horst JP, Loi M et al (2016) Blocking glucocorticoid function in control of salt balance and stress-adaptation. receptors at adolescent age prevents enhanced freezing between Physiol Behav 178:13–20. https​://doi.org/10.1016/j.physb​ repeated cue-exposures after conditioned fear in adult mice eh.2016.12.045 raised under chronic early life stress. Neurobiol Learn Mem de Kloet ER, De Kock S, Schild V, Veldhuis HD (1988) Antiglucocorti- 133:30–38. https​://doi.org/10.1016/j.nlm.2016.05.009 coid RU 38486 attenuates retention of a behaviour and disinhibits Atucha E, Zalachoras I, van den Heuvel JK et al (2015) A mixed glu- the hypothalamic-pituitary adrenal axis at different brain sites. cocorticoid/mineralocorticoid selective modulator with dominant Neuroendocrinology 47:109–115. https://doi.org/10.1159/00012​ ​ antagonism in the male rat brain. Endocrinology 156:4105–4114. 4900 https​://doi.org/10.1210/en.2015-1390 De Kloet ER, Vreugdenhil E, Oitzl MS, Joëls M (1998) Brain corti- Bachmann CG, Linthorst ACE, Holsboer F, Reul JMHM (2003) Effect costeroid receptor balance in health and disease. Endocr Rev of chronic administration of selective glucocorticoid receptor 19:269–301 antagonists on the rat hypothalamic-pituitary-adrenocortical De Bosscher K, Vanden Berghe W, Haegeman G (2003) The interplay axis. Neuropsychopharmacology 28:1056–1067. https​://doi. between the glucocorticoid receptor and nuclear factor-kappaB or org/10.1038/sj.npp.13001​58 activator protein-1: molecular mechanisms for gene repression. Beaudry JL, Dunford EC, Teich T et al (2014) Effects of selective and Endocr Rev 24:488–522 non-selective glucocorticoid receptor II antagonists on rapid- de Kloet ER, Joëls M, Holsboer F (2005) Stress and the brain: from onset diabetes in young rats. PLoS ONE 9:e91248. https​://doi. adaptation to disease. Nat Rev Neurosci 6:463–475. https​://doi. org/10.1371/journ​al.pone.00912​48 org/10.1038/nrn16​83 Belanoff JK, Flores BH, Kalezhan M et al (2001) Rapid reversal of De Bosscher K, Beck IM, Ratman D et al (2016) Activation of the glu- psychotic depression using mifepristone. J Clin Psychophar- cocorticoid receptor in acute inflammation: the SEDIGRAM con- macol 21:516–521. https​://doi.org/10.1097/00004​714-20011​ cept. Trends Pharmacol Sci 37:4–16. https​://doi.org/10.1016/j. 0000-00009​ tips.2015.09.002 Belanoff JK, Rothschild AJ, Cassidy F et al (2002) An open label de Kloet ER, Meijer OC, de Nicola AF et al (2018) Importance of the trial of C-1073 (mifepristone) for psychotic major depression*. brain corticosteroid receptor balance in metaplasticity, cognitive Biol Psychiatr 52:386–392. https​://doi.org/10.1016/S0006​ performance and neuro-inflammation. Front Neuroendocrinol -3223(02)01432​-4 49:124–145. https​://doi.org/10.1016/j.yfrne​.2018.02.003 Blasey CM, Debattista C, Roe R et al (2009) A multisite trial of mife- DeBattista C, Belanoff J (2006) The use of mifepristone in the treat- pristone for the treatment of psychotic depression: a site-by- ment of neuropsychiatric disorders. Trends Endocrinol Metab treatment interaction. Contemp Clin Trials 30:284–288. https​:// 17:117–121. https​://doi.org/10.1016/j.tem.2006.02.006 doi.org/10.1016/j.cct.2009.03.001

1 3 Cellular and Molecular Neurobiology

Deraedt R, Bonnat C, Busigny M et al (1985) Pharmacokinetics of RU induced suppression of hippocampal cell proliferation. PLoS 486. In: The antiprogestin steroid RU 486 and human fertility ONE 7:e46224. https​://doi.org/10.1371/journ​al.pone.00462​24 control. Springer, Boston, pp 103–122 Iglesias T, Verbeeck MAE, De Kloet RE et al (1991) Time- and Deuschle M, Weber B, Colla M et al (1998) Mineralocorticoid receptor region-dependent effect of adrenalectomy on neuropeptide gene also modulates basal activity of hypothalamus-pituitary-adren- expression in rat hippocampus and striatum. Mol Cell Neurosci ocortical system in humans. Neuroendocrinology 68:355–360. 2:485–490. https​://doi.org/10.1016/1044-7431(91)90015​-G https​://doi.org/10.1159/00005​4384 Jacobson L, Akana SF, Cascio CS et al (1989) The adrenocortical sys- Di S, Malcher-Lopes R, Halmos KC, Tasker JG (2003) Nongenomic tem responds slowly to removal of corticosterone in the absence glucocorticoid inhibition via endocannabinoid release in of concurrent stress*. Endocrinology 124:2144–2152. https​:// the hypothalamus: a fast feedback mechanism. J Neurosci doi.org/10.1210/endo-124-5-2144 23:4850–4857 Joëls M (2006) Corticosteroid effects in the brain: U-shape it. Dodt C, Kern W, Fehm L, Born J (1993) Antimineralocorticoid can- Trends Pharmacol Sci 27:244–250. https​://doi.org/10.1016/j. renoate enhances secretory activity of the hypothalamus-pitui- tips.2006.03.007 tary-adrenocortical (HPA) axis in humans. Neuroendocrinology Joëls M, de Kloet ER (1992) Control of neuronal excitability by cor- 58:570–574. https​://doi.org/10.1159/00012​6592 ticosteroid hormones. Trends Neurosci 15:25–30. https​://doi. Ferguson D, Sapolsky R (2007) Mineralocorticoid receptor over- org/10.1016/0166-2236(92)90345​-9 expression differentially modulates specific phases of spatial Joëls M, de Kloet ER (2017) 30 Years of the mineralocorticoid recep- and nonspatial memory. J Neurosci 27:8046–8052. https​://doi. tor: the brain mineralocorticoid receptor: a saga in three episodes. org/10.1523/JNEUR​OSCI.1187-07.2007 J Endocrinol 234:T49–T66. https://doi.org/10.1530/JOE-16-0660​ Fleseriu M, Biller BMK, Findling JW et al (2012) Mifepristone, a glu- Joëls M, Karst H, DeRijk R, de Kloet ER (2008) The coming out of the cocorticoid receptor antagonist, produces clinical and metabolic brain mineralocorticoid receptor. Trends Neurosci 31:1–7. https​ benefits in patients with Cushing’s syndrome. J Clin Endocrinol ://doi.org/10.1016/j.tins.2007.10.005 Metab 97:2039–2049. https​://doi.org/10.1210/jc.2011-3350 Joëls M, Karst H, Sarabdjitsingh RA (2018) The stressed brain of Flores BH, Kenna H, Keller J et al (2006) Clinical and biologi- humans and rodents. Acta Physiol (Oxf) 223:e13066. https​:// cal effects of mifepristone treatment for psychotic depres- doi.org/10.1111/apha.13066​ sion. Neuropsychopharmacology 31:628–636. https​://doi. Karssen AM, Meijer OC, van der Sandt IC et al (2001) Multidrug org/10.1038/sj.npp.13008​84 resistance P-glycoprotein hampers the access of cortisol but not Gaillard RC, Riondel A, Muller AF et al (1984) RU 486: a ster- of corticosterone to mouse and human brain. Endocrinology oid with antiglucocorticosteroid activity that only disinhibits 142:2686–2694. https​://doi.org/10.1210/endo.142.6.8213 the human pituitary-adrenal system at a specific time of day. Karssen AM, Meijer OC, van der Sandt ICJ et al (2002) The role of Proc Natl Acad Sci 81:3879–3882. https​://doi.org/10.1073/ the efflux transporter P-glycoprotein in brain penetration of pred- pnas.81.12.3879 nisolone. J Endocrinol 175:251–260. https​://doi.org/10.1677/ Garner B, Phillips LJ, Bendall S, Hetrick SE (2016) Antiglucocorticoid joe.0.17502​51 and related treatments for psychosis. Cochrane Database Syst Karssen AM, Meijer O, Pons D, de Kloet ER (2004) Localization of Rev https​://doi.org/10.1002/14651​858.CD006​995.pub2 mRNA expression of p-glycoprotein at the blood-brain barrier Groeneweg FL, Karst H, de Kloet ER, Joëls M (2012) Mineralocorti- and in the hippocampus. Ann N Y Acad Sci 1032:308–311. https​ coid and glucocorticoid receptors at the neuronal membrane, reg- ://doi.org/10.1196/annal​s.1314.048 ulators of nongenomic corticosteroid signalling. Mol Cell Endo- Karssen AM, Meijer OC, Berry A et al (2005) Low doses of dexameth- crinol 350:299–309. https​://doi.org/10.1016/j.mce.2011.06.020 asone can produce a hypocorticosteroid state in the brain. Endo- Gruol DJ, Zee MC, Trotter J, Bourgeois S (1994) Reversal of multidrug crinology 146:5587–5595. https://doi.org/10.1210/en.2005-0501​ resistance by RU 486. Cancer Res 54:3088–3091 Karst H, Joëls M (2003) Effect of chronic stress on synaptic currents in Harris AP, Holmes MC, de Kloet ER et al (2013) Mineralocorticoid rat hippocampal dentate gyrus neurons. J Neurophysiol 89:625– and glucocorticoid receptor balance in control of HPA axis and 633. https​://doi.org/10.1152/jn.00691​.2002 behaviour. Psychoneuroendocrinology 38:648–658. https​://doi. Karst H, Berger S, Turiault M et al (2005) Mineralocorticoid receptors org/10.1016/j.psyne​uen.2012.08.007 are indispensable for nongenomic modulation of hippocampal Havel PJ, Busch BL, Curry DL et al (1996) Predominately glucocorti- glutamate transmission by corticosterone. Proc Natl Acad Sci coid agonist actions of RU-486 in young specific-pathogen-free 102:19204–19207. https​://doi.org/10.1073/pnas.05075​72102​ Zucker rats. Am J Physiol 271:R710–R717 Karst H, Berger S, Erdmann G et al (2010) Metaplasticity of amygdalar Heck S, Kullmann M, Gast A et al (1994) A distinct modulating responses to the stress hormone corticosterone. Proc Natl Acad domain in glucocorticoid receptor monomers in the repression of Sci U S A 107:14449–14454. https://doi.org/10.1073/pnas.09143​ ​ activity of the transcription factor AP-1. EMBO J 13:4087–4095. 81107​ https​://doi.org/10.1002/j.1460-2075.1994.tb067​26.x Kolber BJ, Wieczorek L, Muglia LJ (2008) Hypothalamic-pituitary- Heikinheimo O, Kontula K, Croxatto H et al (1987) Plasma con- adrenal axis dysregulation and behavioral analysis of mouse centrations and receptor binding of RU 486 and its metabo- mutants with altered glucocorticoid or mineralocorticoid recep- lites in humans. J Steroid Biochem 26:279–284. https​://doi. tor function. Stress 11:321–338. https​://doi.org/10.1080/10253​ org/10.1016/0022-4731(87)90083​-5 89070​18210​81 Hill MN, Tasker JG (2012) Endocannabinoid signaling, glucocorti- Kroon J, Koorneef LL, van den Heuvel JK et al (2018) Selective glu- coid-mediated negative feedback, and regulation of the hypotha- cocorticoid receptor antagonist CORT125281 activates brown lamic-pituitary-adrenal axis. Neuroscience 204:5–16. https://doi.​ adipose tissue and alters lipid distribution in male mice. Endo- org/10.1016/j.neuro​scien​ce.2011.12.030 crinology 159:535–546. https​://doi.org/10.1210/en.2017-00512​ Hollis F, van der Kooij MA, Zanoletti O et al (2015) Mitochondrial Lachize S, Apostolakis EM, van der Laan S et al (2009) Steroid recep- function in the brain links anxiety with social subordination. tor coactivator-1 is necessary for regulation of corticotropin- Proc Natl Acad Sci 112:15486–15491. https​://doi.org/10.1073/ releasing hormone by chronic stress and glucocorticoids. Proc pnas.15126​53112​ Natl Acad Sci USA 106:8038–8042. https​://doi.org/10.1073/ Hu P, Oomen C, van Dam A-M et al (2012) a single-day treatment with pnas.08120​62106​ mifepristone is sufficient to normalize chronic glucocorticoid

1 3 Cellular and Molecular Neurobiology

Lähteenmäki P, Heikinheimo O, Croxatto H et al (1987) Pharmacoki- Murphy BE, Filipini D, Ghadirian AM (1993) Possible use of gluco- netics and metabolism of RU 486. J Steroid Biochem 27:859–863 corticoid receptor antagonists in the treatment of major depres- Lai M, Horsburgh K, Bae SE et al (2007) Forebrain mineralocorticoid sion: preliminary results using RU 486. J Psychiatry Neurosci receptor overexpression enhances memory, reduces anxiety and 18:209–213 attenuates neuronal loss in cerebral ischaemia. Eur J Neurosci Nguyen ET, Streicher J, Berman S et al (2017) A mixed gluco- 25:1832–1842. https://doi.org/10.1111/j.1460-9568.2007.05427​ ​ corticoid/mineralocorticoid receptor modulator dampens .x endocrine and hippocampal stress responsivity in male rats. Lecureur V, Fardel O, Guillouzo A (1994) The antiprogestatin drug RU Physiol Behav 178:82–92. https​://doi.org/10.1016/j.physb​ 486 potentiates doxorubicin cytotoxicity in multidrug resistant eh.2017.01.020 cells through inhibition of P-glycoprotein function. FEBS Lett Nguyen ET, Caldwell JL, Streicher J et al (2018) Differential effects 355:187–191. https​://doi.org/10.1016/0014-5793(94)01186​-9 of imipramine and CORT118335 (glucocorticoid receptor mod- Loi M, Sarabdjitsingh RA, Tsouli A et al (2017) Transient prepu- ulator/mineralocorticoid receptor antagonist) on brain-endo- bertal mifepristone treatment normalizes deficits in contextual crine stress responses and depression-like behavior in female memory and neuronal activity of adult male rats exposed to rats. Behav Brain Res 336:99–110. https​://doi.org/10.1016/j. maternal deprivation. Eneuro. https​://doi.org/10.1523/ENEUR​ bbr.2017.08.045 O.0253-17.2017 Oitzl MS, de Kloet ER (1992) Selective corticosteroid antagonists Mason BL, Pariante CM, Thomas SA (2012) Changes in the brain modulate specific aspects of spatial orientation learning. Behav accumulation of glucocorticoids in abcb1a-deficient CF-1 Neurosci 106:62–71. https://doi.org/10.1037/0735-7044.106.1.62​ mice. J Neuroendocrinol 24:1440–1446. https​://doi.org/10.111 Oitzl MS, Fluttert M, Ron de Kloet E (1994) The effect of corticoster- 1/j.1365-2826.2012.02353​.x one on reactivity to spatial novelty is mediated by central miner- Matosin N, Halldorsdottir T, Binder EB (2018) Understanding the alocorticosteroid receptors. Eur J Neurosci 6:1072–1079. https​ molecular mechanisms underpinning gene by environment inter- ://doi.org/10.1111/j.1460-9568.1994.tb006​04.x actions in psychiatric disorders: the FKBP5 model. Biol Psychi- Oitzl MS, van Haarst AD, Sutanto W, de Kloet ER (1995) Corticoster- atr 83:821–830. https​://doi.org/10.1016/j.biops​ych.2018.01.021 one, brain mineralocorticoid receptors (MRs) and the activity of Mayer JL, Klumpers L, Maslam S et al (2006) Brief treatment with the hypothalamic-pituitary-adrenal (HPA) axis: the Lewis rat as the glucocorticoid receptor antagonist mifepristone normal- an example of increased central MR capacity and a hyporespon- ises the corticosterone-induced reduction of adult hippocampal sive HPA axis. Psychoneuroendocrinology 20:655–675. https​:// neurogenesis. J Neuroendocrinol 18:629–631. https​://doi.org/ doi.org/10.1016/0306-4530(95)00003​-7 10.1111/j.1365-2826.2006.01455​.x Oitzl MS, van Haarst AD, de Kloet ER (1997) Behavioral and neuroen- Meijer OC, De Lange ECM, Breimer DD et al (1998) Penetration of docrine responses controlled by the concerted action of central dexamethasone into brain glucocorticoid targets is enhanced mineralocorticoid (MRS) and glucocorticoid receptors (GRS). in mdr1A P-glycoprotein knockout mice. Endocrinology Psychoneuroendocrinology 22(Suppl 1):S87–S93. https​://doi. 139:1789–1793 org/10.1016/S0306​-4530(97)00020​-6 Meijer OC, Steenbergen PJ, De Kloet ER (2000) Differential expres- Oitzl MS, Fluttert M, Sutanto W, de Kloet ER (1998) Continuous sion and regional distribution of steroid receptor coactiva- blockade of brain glucocorticoid receptors facilitates spatial tors SRC-1 and SRC-2 in brain and pituitary. Endocrinology learning and memory in rats. Eur J Neurosci 10:3759–3766. https​ 141:2192–2199. https​://doi.org/10.1210/endo.141.6.7489 ://doi.org/10.1046/j.1460-9568.1998.00381​.x Meijer OC, Kalkhoven E, van der Laan S et al (2005) Steroid recep- Oitzl MS, Reichardt HM, Joëls M, de Kloet ER (2001) Point mutation tor coactivator-1 splice variants differentially affect corticos- in the mouse glucocorticoid receptor preventing DNA binding teroid receptor signaling. Endocrinology 146:1438–1448. https​ impairs spatial memory. Proc Natl Acad Sci U S A 98:12790– ://doi.org/10.1210/en.2004-0411 12795. https​://doi.org/10.1073/pnas.23131​3998 Meijer OC, Koorneef LL, Kroon J (2018) Glucocorticoid receptor Oomen CA, Mayer JL, de Kloet ER et al (2007) Brief treatment with modulators. Ann Endocrinol (Paris) 79:107–111. https​://doi. the glucocorticoid receptor antagonist mifepristone normalizes org/10.1016/j.ando.2018.03.004 the reduction in neurogenesis after chronic stress. Eur J Neurosci Meyer M, Gonzalez Deniselle MC, Hunt H et al (2014) The selec- 26:3395–3401. https://doi.org/10.1111/j.1460-9568.2007.05972​ ​ tive glucocorticoid receptor modulator CORT108297 restores .x faulty hippocampal parameters in Wobbler and corticosterone- Picard M, Juster R-P, McEwen BS (2014) Mitochondrial allostatic treated mice. J Steroid Biochem Mol Biol 143:40–48. https​:// load puts the ‘gluc’ back in glucocorticoids. Nat Rev Endocrinol doi.org/10.1016/j.jsbmb​.2014.02.007 10:303–310. https​://doi.org/10.1038/nrend​o.2014.22 Meyer M, Lara A, Hunt H et al (2018) The selective glucocorticoid Pineau F, Canet G, Desrumaux C et al (2016) New selective glucocor- receptor modulator cort 113176 reduces neurodegeneration ticoid receptor modulators reverse amyloid-β peptide-induced and neuroinflammation in wobbler mice spinal cord. Neuro- hippocampus toxicity. Neurobiol Aging 45:109–122 science 384:384–396. https​://doi.org/10.1016/j.neuro​scien​ Ratka A, Sutanto W, De Kloet R (1988) Long-lasting glucocorticoid ce.2018.05.042 suppression of opioid-induced antinociception. Neuroendocrinol- Mifsud KR, Reul JMHM (2016) Acute stress enhances heterodimeri- ogy 48:439–444. https​://doi.org/10.1159/00012​5046 zation and binding of corticosteroid receptors at glucocorticoid Ratka A, Sutanto W, Bloemers M, de Kloet ER (1989) On the role target genes in the hippocampus. Proc Natl Acad Sci U S A of brain mineralocorticoid (type I) and glucocorticoid (type II) 113:11336–11341. https​://doi.org/10.1073/pnas.16052​46113​ receptors in neuroendocrine regulation. Neuroendocrinology Moraitis AG, Block T, Nguyen D, Belanoff JK (2017) The role of 50:117–123. https​://doi.org/10.1159/00012​5210 glucocorticoid receptors in metabolic syndrome and psychiat- Reichardt HM, Kaestner KH, Tuckermann J et al (1998) DNA binding ric illness. J Steroid Biochem Mol Biol 165:114–120. https​:// of the glucocorticoid receptor is not essential for survival. Cell doi.org/10.1016/j.jsbmb​.2016.03.023 93:531–541. https​://doi.org/10.1016/S0092​-8674(00)81183​-6 Munck A, Guyre PM, Holbrook NJ (1984) Physiological functions Repunte-Canonigo V, Shin W, Vendruscolo LF et al (2015) Identifying of glucocorticoids in stress and their relation to pharmaco- candidate drivers of alcohol dependence-induced excessive drink- logical actions. Endocr Rev 5:25–44. https​://doi.org/10.1210/ ing by assembly and interrogation of brain-specific regulatory edrv-5-1-25

1 3 Cellular and Molecular Neurobiology

networks. Genome Biol 16:68. https​://doi.org/10.1186/s1305​ Teich T, Dunford EC, Porras DP et al (2016) Glucocorticoid antago- 9-015-0593-5 nism limits adiposity rebound and glucose intolerance in young Reul JM, de Kloet ER (1985) Two receptor systems for corticos- male rats following the cessation of daily exercise and caloric terone in rat brain: microdistribution and differential occupa- restriction. Am J Physiol Metab 311:E56–E68. https​://doi. tion. Endocrinology 117:2505–2511. https​://doi.org/10.1210/ org/10.1152/ajpen​do.00490​.2015 endo-117-6-2505 Ulrich-Lai YM, Herman JP (2009) Neural regulation of endocrine and Reul JM, de Kloet ER, van Sluijs FJ et al (1990) Binding character- autonomic stress responses. Nat Rev Neurosci 6:397–409. https​ istics of mineralocorticoid and glucocorticoid receptors in dog ://doi.org/10.1038/nrn26​47 brain and pituitary. Endocrinology 127:907–915. https​://doi. Thomson F, Craighead M (2008) Innovative approaches for the treat- org/10.1210/endo-127-2-907 ment of depression: targeting the HPA axis. Neurochem Res Revsin Y, Rekers NV, Louwe MC et al (2009) Glucocorticoid recep- 33:691–707. https​://doi.org/10.1007/s1106​4-007-9518-3 tor blockade normalizes hippocampal alterations and cognitive van Gemert NG, Joëls M (2006) Effect of chronic stress and mife- impairment in streptozotocin-induced type 1 diabetes mice. Neu- pristone treatment on voltage-dependent Ca 2 + currents in rat ropsychopharmacology 34:747–758. https​://doi.org/10.1038/ hippocampal dentate gyrus. J Neuroendocrinol 18:732–741. https​ npp.2008.136 ://doi.org/10.1111/j.1365-2826.2006.01472​.x Ronacher K, Hadley K, Avenant C et al (2009) Ligand-selective trans- van Dijk AM, van Wimersma Greidanus TB, Burbach JP et al (1981) activation and transrepression via the glucocorticoid receptor: Brain adrenocorticotrophin after adrenalectomy and sham-oper- Role of cofactor interaction. Mol Cell Endocrinol 299:219–231. ation of rats. J Endocrinol 88:243–253. https​://doi.org/10.1677/ https​://doi.org/10.1016/j.mce.2008.10.008 joe.0.08802​43 Rozeboom AM, Akil H, Seasholtz AF (2007) Mineralocorticoid recep- van Eekelen JAM, Kiss JZ, Westphal HM, de Kloet ER (1987) Immu- tor overexpression in forebrain decreases anxiety-like behav- nocytochemical study on the intracellular localization of the type ior and alters the stress response in mice. Proc Natl Acad Sci 2 glucocorticoid receptor in the rat brain. Brain Res 436:120– 104:4688–4693. https​://doi.org/10.1073/pnas.06060​67104​ 128. https​://doi.org/10.1016/0006-8993(87)91563​-0 Sapolsky RM, Romero LM, Munck AU (2000) How do glucocorticoids Van Haarst AD, Oitzl MS, Workel JO, De Kloet ER (1996) Chronic influence stress responses? Integrating permissive, suppressive, brain glucocorticoid receptor blockade enhances the rise in circa- stimulatory, and preparative actions. Endocr Rev 21:55–89. https​ dian and stress-induced pituitary-adrenal activity. Endocrinology ://doi.org/10.1210/edrv.21.1.0389 137:4935–4943. https​://doi.org/10.1210/en.137.11.4935 Sarrieau A, Dussaillant M, Agid F et al (1986) Autoradiographic locali- van Haarst AD, Oitzl MS, de Kloet ER (1997) Facilitation of feed- zation of glucocorticoid and progesterone binding sites in the back inhibition through blockade of glucocorticoid receptors in human post-mortem brain. J Steroid Biochem 25:717–721. https​ the hippocampus. Neurochem Res 22:1323–1328. https​://doi. ://doi.org/10.1016/0022-4731(86)90300​-6 org/10.1023/A:10220​10904​600 Schinkel AH, Wagenaar E, van Deemter L et al (1995) Absence of Van Weert LTCM, Buurstede JC, Mahfouz A et al (2017) NeuroD fac- the mdr1a P-Glycoprotein in mice affects tissue distribution and tors discriminate mineralocorticoid from glucocorticoid receptor pharmacokinetics of dexamethasone, digoxin, and cyclosporin A. DNA binding in the male rat brain. Endocrinology 158:1511– J Clin Invest 96:1698–1705. https​://doi.org/10.1172/JCI11​8214 1522. https​://doi.org/10.1210/en.2016-1422 Schmidt MV, Sterlemann V, Wagner K et al (2009) Postnatal glucocor- van den Heuvel JK, Boon MR, van Hengel I et al (2016) Identification ticoid excess due to pituitary glucocorticoid receptor deficiency: of a selective glucocorticoid receptor modulator that prevents Differential short- and long-term consequences. Endocrinology both diet-induced obesity and inflammation. Br J Pharmacol 150:2709–2716. https​://doi.org/10.1210/en.2008-1211 173:1793–1804. https​://doi.org/10.1111/bph.13477​ Schwabe L, Schächinger H, de Kloet ER, Oitzl MS (2010) Corticoster- van der Hoeven RA, Hofte AJ, Frenay M et al (1997) Liquid chroma- oids operate as a switch between memory systems. J Cogn Neu- tography-mass spectrometry with on-line solid-phase extraction rosci 22:1362–1372. https​://doi.org/10.1162/jocn.2009.21278​ by a restricted-access C18 precolumn for direct plasma and urine Schwabe L, Tegenthoff M, Höffken O, Wolf OT (2013) Mineralocor- injection. J Chromatogr A 762:193–200. https://doi.org/10.1016/​ ticoid receptor blockade prevents stress-induced modulation of S0021​-9673(96)01004​-7 multiple memory systems in the human brain. Biol Psychiatry van der Veen R, Boshuizen MCS, de Kloet ER (2013) Mifepristone 74:801–808. https​://doi.org/10.1016/j.biops​ych.2013.06.001 treatment affects the response to repeated amphetamine injec- Solomon MB, Wulsin AC, Rice T et al (2014) The selective gluco- tions, but does not attenuate the expression of sensitization. Psy- corticoid receptor antagonist CORT 108297 decreases neu- chopharmacology 230:547–556. https​://doi.org/10.1007/s0021​ roendocrine stress responses and immobility in the forced swim 3-013-3176-8 test. Horm Behav 65:363–371. https​://doi.org/10.1016/j.yhbeh​ Vendruscolo LF, Barbier E, Schlosburg JE et al (2012) Corticosteroid- .2014.02.002 dependent plasticity mediates compulsive alcohol drinking in Sousa N (2016) The dynamics of the stress neuromatrix. Mol Psychia- rats. J Neurosci 32:7563–7571. https​://doi.org/10.1523/JNEUR​ try 21:302–312. https​://doi.org/10.1038/mp.2015.196 OSCI.0069-12.2012 Spencer RL, Kim PJ, Kalman BA, Cole MA (1998) Evidence for min- Vendruscolo LF, Estey D, Goodell V et al (2015) Glucocorticoid eralocorticoid receptor facilitation of glucocorticoid receptor- receptor antagonism decreases alcohol seeking in alcohol- dependent regulation of hypothalamic-pituitary-adrenal axis dependent individuals. J Clin Invest 125:3193–3197. https://doi.​ activity. Endocrinology 139:2718–2726. https://doi.org/10.1210/​ org/10.1172/JCI79​828 en.139.6.2718 Wulsin AC, Herman JP, Solomon MB (2010) Mifepristone decreases Spiga F, Knight DM, Droste SK et al (2011) Differential effect of depression-like behavior and modulates neuroendocrine and cen- glucocorticoid receptor antagonists on glucocorticoid receptor tral hypothalamic-pituitary-adrenocortical axis responsiveness nuclear translocation and DNA binding. J Psychopharmacol to stress. Psychoneuroendocrinology 35:1100–1112. https://doi.​ 25:211–221. https​://doi.org/10.1177/02698​81109​34817​5 org/10.1016/j.psyne​uen.2010.01.011 Stranahan AM, Arumugam TV, Cutler RG et al (2008) Diabetes Young EA, Lopez JF, Murphy-Weinberg V et al (1998) The role of impairs hippocampal function through glucocorticoid-mediated mineralocorticoid receptors in hypothalamic-pituitary-adrenal effects on new and mature neurons. Nat Neurosci 11:309–317. axis regulation in humans 1. J Clin Endocrinol Metab 83:3339– https​://doi.org/10.1038/nn205​5 3345. https​://doi.org/10.1210/jcem.83.9.5077

1 3 Cellular and Molecular Neurobiology

Zalachoras I, Houtman R, Atucha E et al (2013) Differential targeting Zalachoras I, Verhoeve SL, Toonen LJ et al (2016) Isoform switch- of brain stress circuits with a selective glucocorticoid receptor ing of steroid receptor co-activator-1 attenuates glucocorticoid- modulator. Proc Natl Acad Sci USA 110:7910–7915. https://doi.​ induced anxiogenic amygdala CRH expression. Mol Psychiatr org/10.1073/pnas.12194​11110​ 21:1733–1739. https​://doi.org/10.1038/mp.2016.16

1 3